| |  | Mar 18, 2026 | | | | | Supported by |  | | | | | | Happy Wednesday! Microsoft overhauls AI leadership. Amazon CEO says AWS could reach $600 billion in sales by 2036. Nvidia restarts manufacturing of the H200 chips for China.
| | | | Microsoft is overhauling its executive ranks to give CEO Satya Nadella more direct oversight of engineering teams working on its Copilot chatbots, Nadella said in a memo to staff Tuesday. The reorganization will bring together engineering divisions working on the consumer version of Copilot and the version geared towards business customers that Microsoft sells as part of its Office 365 suite, which were previously housed in different organizations, Nadella said. As part of the move, corporate vice president Jacob Andreou, who previously reported to consumer AI leader Mustafa Suleyman, will be promoted to executive vice president of Copilot, reporting directly to Nadella. Meanwhile, Suleyman's organization will now focus primarily on training frontier AI models, which was already part of its mandate alongside building the consumer Copilot app, Nadella said. The reorganization comes a week after Nadella told staff that he would reorganize some of the teams working on enterprise Copilots following Office and Windows boss Rajesh Jha's decision to retire. Alongside Andreou and Suleyman, business Copilot leaders Charles Lamanna, Ryan Roslansky, Perry Clarke will form a "Copilot leadership team" that Nadella will oversee directly, he said Tuesday. The leadership shakeup comes as Microsoft races to make its Copilot products more indispensable to both consumers and enterprises as it faces competition from AI labs like Anthropic and OpenAI. The company recently announced new Copilot features powered by Anthropic's models meant to automate more complex business tasks, similar to Anthropic's Cowork product. | | | | Amazon CEO Andy Jassy, in an all-hands meeting with employees, said he believes customer demand for AI could propel Amazon Web Services to a $600 billion annualized sales pace within the next decade, up from around $130 billion in total sales in 2025, Reuters reported. Jassy's bullishness is likely fueled in part by the cloud giant's growing relationship with OpenAI, in which the startup is committing to spend nearly $140 billion on its cloud servers over the next eight years. OpenAI and Amazon are developing a new cloud service to help AWS customers build custom AI agents— powered by OpenAI technology—to automate business tasks. Amazon also has a close relationship with Anthropic, whose annualized revenue pace is quickly catching up to that of OpenAI, and growing faster. While Amazon got off to a slow start in AI, giving an opening to Microsoft and Google, working closely with the top two AI labs could cushion the blow as more of its customers start building AI applications. OpenAI also recently inked a deal with AWS to sell its AI to U.S. government employees for both classified and unclassified work, which will help OpenAI to support the Pentagon as part of its recent agreement with the agency. | | | | Nvidia is in the process of restarting manufacturing for the H200 chips that it plans to sell to Chinese customers, Nvidia CEO Jensen Huang said at a press conference Tuesday. Huang said Nvidia had received purchase orders from customers in China, implying that the Chinese government has given local businesses the green light to buy the chips. Nvidia's ability to sell the H200 chips to China has been up in the air for several months, both because of restrictions imposed by the U.S. government and Chinese authorities. The Trump administration recently signaled it would relax its restrictions but the Chinese government's approval remained uncertain. Nvidia disclosed in February that it received government approvals for small amounts of H200 sales for Chinese customers, but that it did not know whether the Chinese government would allow the sales. Executives did not project any revenue from Chinese customers in its forecast. At the press conference, Huang said "President Trump's intention is that the United States should have a leadership position in access to Nvidia's best technology. However, he would like us to compete worldwide and not concede those markets unnecessarily." | | | | The Chinese government plans to penalize people linked to Meta Platforms' $2 billion acquisition of Manus, a Singapore-headquartered AI application startup that was originally founded in China, The New York Times reported. China's National Development and Reform Commission, a powerful agency responsible for economic planning, held a meeting last week with executives of Meta and Manus to discuss Beijing's concerns about the deal, according to the Times. In January, The Financial Times reported that the Chinese government was reviewing the acquisition, which was announced by Manus in December, due to potential violations of China's technology export control. The Times said the Chinese government's penalties could include exit bans that would prevent Manus executives who enter China from leaving the country. Any actions by Beijing on the Manus deal could have an impact on many Chinese tech entrepreneurs who are building AI products for the global market and moving their businesses outside China. Many Chinese founders launching their AI apps in Western markets are positioning their startups as global companies headquartered in Singapore to avoid potential political scrutiny in the U.S. that Chinese firms face. "The transaction complied fully with applicable law," Meta said in a statement to the Times. "The outstanding team at Manus is now deeply integrated into Meta." | | | | Nvidia CEO Jensen Huang said his firm has invested billions of dollars lately in upstart cloud providers that buy and rent out Nvidia chips because traditional cloud providers won't be able to satisfy rising demand from large AI developers. "The reason why we know they're going to be home runs is because we already see the pipeline of businesses that are coming toward it," Huang said during a press conference at Nvidia's annual GTC conference in San Jose, Calif. Investors have asked questions about Nvidia's circular transactions with the young cloud firms, known as neoclouds, including CoreWeave, Lambda, Nscale and Nebius. Nvidia has given financial guarantees to CoreWeave, whose stock has doubled since going public last year. Many of the other neoclouds are preparing to go public as well. Huang said he views the investment risk in these firms as "extremely low." He said he believes data center and cloud firms could raise money without Nvidia's help. "We get involved as an extra booster, but it's really a fantastic way to be part of the ecosystem," he said. | | | | Chinese AI model developer Zhipu is planning to bring all of its AI product teams, including those working on various apps as well as hardware like AI glasses, under a new unit in a major reorganization, according to two people with direct knowledge of the matter. The creation of the new unit, which will report directly to Zhipu cofounder Tang Jie, is part of the company's effort to strengthen its product development for consumers and enterprise customers, the people said. Zhipu, which went public in Hong Kong in January under the corporate name Knowledge Atlas Technology, is trying to create new sources of revenue by rolling out more competitive products. Beijing-based Zhipu, founded in 2019 by a group of prominent AI researchers at Tsinghua University, is known primarily for its open-source large language models. The company's latest LLM released last month, GLM-5, has received positive reviews for its ability to power AI agents. That model is currently ranked No. 3 in terms of agentic capabilities on an AI model leaderboard compiled by Artificial Analysis, behind OpenAI's GPT-5.4 and Anthropic's Claude Opus 4.6. According to its IPO prospectus, Zhipu derived 85% of its revenue in the first half of 2025 from helping big Chinese enterprise customers, including state-owned firms, use its AI models in their facilities. As revenue from such AI deployment projects may not be recurring, Zhipu needs to step up its effort to build products that can generate more revenue, one of the people said. | | | | Mastercard said it agreed to buy BVNK, a stablecoin payment startup, for up to $1.8 billion, including $300 million in contingent payments, in its biggest bet that more companies will use the cryptocurrency to move money globally. With the acquisition, Mastercard is defending against the threat that stablecoins may disrupt traditional methods of payments such as cards. The deal allows Mastercard to have the technology, licenses and bank partners to offer stablecoin payments to its clients. BVNK, founded in 2021 and based in London, allows businesses to settle and process payments using stablecoins by providing the conversion between local currency and stablecoins. It counts human resources startup Deel and payments company Worldpay among its clients and processed $30 billion in annualized volume in 2025. BVNK is a competitor to Bridge, the stablecoin startup acquired by Stripe last year in a $1.1 billion deal. BVNK was backed by Visa, Coinbase Ventures, Haun Ventures, and Citi Ventures, and raised $50 million at a reported $750 million valuation in late 2024. | | | | Nvidia is working on a version of its Groq AI chips that can be sold in China, according to Reuters, citing two people familiar with the matter, as the company works to regain ground in a market where U.S. export controls have restricted its presence. Nvidia expects these new chips to ship in May, Reuters reported. The move comes as CEO Jensen Huang said earlier this week that Nvidia has restarted manufacturing of the H200, its previous flagship chip, after securing export licenses from the U.S. government. Nvidia signed a $17 billion licensing deal with Groq late last year, and the chips it gained from that partnership are designed mainly for inference, the process by which an AI model responds to a user prompt. The new China-bound chips are a modified variant of the Groq design and will be Nvidia's latest offering to the Chinese market, one step up from the older-generation H200. But Beijing's drive for technological self-sufficiency has pushed domestic tech firms to cut their reliance on foreign chip suppliers. Local tech giants such as Huawei Technologies have spent years developing their own inference chips, building a market presence that competes against Nvidia's offerings. | | | | AI startup You.com, valued at $1.5 billion after a funding round in September, is changing its senior leadership as it focuses on helping businesses adopt AI. Bryan McCann, You.com's CTO and co-founder, recently left to join Anthropic as a member of technical staff, according to his LinkedIn profile. Saahil Jain, one of the startup's early engineering hires, has been promoted to CTO, said Richard Socher, CEO and co-founder of You.com, in a blog post. The CTO change comes as You.com, which started in consumer search, is increasingly targeting business customers with tools that help them analyze documents and create reports. That move has put You.com into more direct competition with larger AI startups like OpenAI and Databricks. Socher, a former chief scientist at Salesforce, is aiming to raise hundreds of millions of dollars for a new AI research startup he will oversee, according to a person with direct knowledge of the matter. Socher said at the time that he will remain as CEO of You.com. | | | | | Popular articles By Anissa Gardizy, Valida Pau and Stephanie Palazzolo
By Julia Hornstein, Anissa Gardizy and Sri Muppidi
By Sri Muppidi and Aaron Holmes
| | | | | Opportunities Empower your teams to stay ahead of market trends with the most trusted tech journalism.
Learn more
Reach The Information's influential audience with your message. Connect with our team | | | | | |
0 comentários:
Postar um comentário